References

advertisement
Some remarks on non-conceptual word meaning and truth-conditional
content in Robyn Carston’s pragmatics
1. Introduction
[…] an account of word meanings as non-conceptual (semantically underspecified)
would be the completing component of this view of the relation between language and
thought: not only do sentence meanings underdetermine thought, but the basic
constituents of sentences (words) underdetermine the basic constituents of thoughts
(concepts). If this account turns out to be right, it’s not just that we don’t always say
what we intend our hearers to take us to mean but that it is simply not possible to say
what we mean. (Carston, 2013, 203)
Over the last two decades Relevance Theory (Sperber and Wilson, 1986[1995]; Wilson
and Sperber, 2004) has undergone fundamental changes.1 Robyn Carston’s contribution has
been substantial. Her latest proposals (Carston, 2013, 2012) focusing on the interpretation of
standing word meaning2 seem to lead to even more revolutionary reconstruction of the theory.
There are two goals to be achieved in this paper. The first one is expository: to show
how stable, conventionalized word meaning has been interpreted in Relevance Theory, and,
more specifically, whether and, if so, how word meaning contributes to the truth-conditional
content of explicatures.
The second goal is discursive. It seems that Carston’s view that word meaning is nonconceptual and non-semantic but schematic and not determinate is not compatible with the
basic assumptions of Relevance Theory because it requires a redefinition of logical form,
explicature and logical form development. It also needs clarification of the nature of
1
For example, from syntax-semantics-pragmatics to syntax-pragmatics distinction, from linguistic semantics to
lexical pragmatics, from concepts understood as stable elementary mental structures to occasion-specific ad hoc
concepts and many more.
2
More precisely, of nouns, verbs and adjectives meaning. Referring to word meaning in the paper is restricted to
the meaning of these three open-class parts of speech.
schematic word meaning, of concepts constituting explicature and of the relation between the
two.
2. From concepts to concept schemas
The canonical RT view on word meaning is similar to the minimalist semantic position3
(Carston, 2013, 196), according to which word meanings are concepts and concepts have a
referential semantics (Carston, 2012; Wilson and Carston, 2007).
Wilson (2003, 274) writes that “I will adopt a simple model of linguistic semantics4 that
treats words as encoding mentally-represented concepts […], which constitute their linguistic
meaning and determine what might be called their linguistically specified denotation.”
(Carston, 2013, 196)
Within this broad view (Sperber and Wilson, 2008), “majority of open-class words
encode concepts, which have a denotation, an externalist semantics, and so contribute directly
to truth-conditional content.” (Carston, 2013, 195-196)
It follows (Carston, 2013, 178, 180) that words encode concepts, constituting a
minimalist mental lexicon: mappings from lexical items to simple concepts.
Concepts are mental representational entities of the Fodorian (1998) type5 (Carston,
2013, 185). They are basic constituents of our language of thought, each of which is a storage
point for encyclopaedic information. These are their two functions in our cognitive life.
Concepts have a referential semantics, that is, they have a denotation: they refer to entities,
properties and activities in the mind-external world.
Within RT lexical pragmatics (Wilson, 2003; Wilson and Carston, 2007; Sperber and
Wilson, 2008), word meaning is pragmatically inferred on top of being encoded. “All words
3
See page 6.
Originally in RT (Sperber an Wilson 1986[1995]; Carston, 1991) a certain division of labour was assumed
between linguistic semantics and truth-conditional semantics. With the emergence of the RT lexical pragmatics,
the role of the former has gradually disappeared.
5
Full-fledged, non-decompositional atomic concepts with no content-construction inferential connection or
knowledge structures.
4
behave as if they encoded pro-concepts, that is […] the concept it is used to convey in a given
utterance has to be contextually worked out.” (Sperber and Wilson 1998, 195[2012]) As a
result of this pragmatic inference, ad hoc concepts may be constructed (or retrieved). They are
occasion-specific and understood in denotational terms (their denotation is narrower and/or
broader than the denotation of the concept encoded by a word triggering ad hoc concept
construction).6
Within RT lexical pragmatics “the pragmatic process of inferring the intended content is
optional because ‘it may so happen that the intended concept is the very one encoded by the
word’ (Sperber and Wilson, 1998, 197[2012])” (Carston, 2013, 185)
At this state of word meaning interpretation “the overall picture of (encoded) word
meaning in RT is quite heterogeneous: procedures of various sorts, pro-concepts and concepts
(denotational entities) all of which provide input to the pragmatic interpretation process, and
all of which, concepts included, function merely as clues or evidence guiding and
constraining processes of pragmatic inference whose goal is the recovery of the intended
interpretation of the utterance.” (Carston, 2013, 196)
The main theses of Robyn Carston’s (2013, 2012) pragmatics, concerning the
stable/standing meaning of three open-class items, that is nouns, verbs and adjectives are as
follows:
a) Stable (lexicalized) word meanings are non-conceptual, that is non-semantic (Carston,
2013, 184). Words do not encode full-fledged concepts; in fact, there is no literal,
encoded meaning of lexical items (Carston, 2013, 200). As in RT lexical pragmatics,
words are merely pointers to or evidence for a speaker’s meaning.
b) “Each word comes with its own distinct but schematic meaning, which functions as a
constraint on the general pragmatic process of accessing or constructing a concept, a
6
For a critical view on RT ad hoc concepts see, among others, Mioduszewska (in press).
process which is wholly motivated by the goal of the pragmatic system which is to
deliver speaker meaning.” (Carston, 2013, 200) The problem is what would be the
lexical concept that comprises the stable meaning of a word whose use gives rise to a
range of distinct ad hoc concepts. For example, is there any definite thought at all that
engages the general concept encoded by the verb open (Carston, 2012, 2013, 185),
covering all its distinct senses (open the door, open one’s mouth, open a meeting…)?
Perhaps it is better “to move to a different non-conceptual (non-semantic) view on the
nature of standing word meaning.” (Carston, 2013, 185)
c) In view of the schematic, non-conceptual stable meaning of open-class words, the
pragmatic inferential process of constructing (contextual) ad-hoc concepts becomes
obligatory (Carston 2013, 187). This process allows for only one of a range of
concepts to be the first one accessed or constructed, as determined by considerations
of relevance.
d) Word meanings are conceptually underspecified (Carston, 2013, 184, 187, 196, 197,
201).7
e) Meaning ≠ sense. Open-class items meaning is the underspecified, non-conceptual,
non-semantic standing meaning of a word. Open-class items senses are fully
conceptual semantic entities that words are used to express by individual language
users in their individual utterances. Speakers express fully semantic entities, that is
concepts, components of the thoughts/propositions expressed (Carston, 2013, 201).8
7
The proof of underdeterminacy of standing word meaning comes from an analysis of polysemy (Bosch, 2009)
and its cognitive foundations (Frisson, 2009) endorsed by Carston. Since all open-class words are (potentially)
polysemous (Carston, 2013, 192), their meaning is underspecified, too. For the needs of the paper, the validity of
the reasoning is irrelevant, as what is examined is the (in)compatibility of Carston’s view with the general RT
assumptions.
8
External support for this position comes, for example, from Bosch (2009, 99) in Carston (2013, 185) “Lexical
semantics is much less specified than is often assumed and only contains structural constraints over the kind of
conceptual entities that can be denoted by the lexical item, but does not contain the conceptual content.”
f) This non-conceptual view of word meaning is compatible with RT.9
According to Robyn Carston, her interpretation of standing word meaning has some
cognitive and theoretical advantages. On the processing effort side, “rather than activating one
or more specific senses, readers initially activate a single, semantically underspecified
meaning. This abstract meaning is the same for the established senses of a word, that is, the
same underspecified meaning encompasses all semantically related interpretations of a word
that are known to a reader.” (Frisson, 2009, 116) in (Carston, 2013, 191-192)
On this account, there is no need to explicate the nature and functioning of full-fledged
concepts. On the other hand, an explanation is offered of the fact that bundles of related
senses (polysemy) seem to accrue around a word form (Carston, 2013, 200).
The account faces some problems too. The most important one boils down to the
question how one gets from non-conceptual, non-semantic word meaning to semantic, truthconditional and conceptual word senses (Carston, 2013, 187, 195 note 20). Another concerns
the organization and tagging of encyclopaedic information (Carston, 2013, 180).
A lot more needs to be said about the nature and functioning of postulated nonconceptual, non-truth-conditional but word invariant concept schemas to make them
cognitively and operationally transparent. Similarly, the status and properties of full-fledged
mental concepts cannot be left unexplained because, although they are no longer needed as
word meanings, they remain constitutive components of speakers’ thoughts (propositions
expressed), which are the meaning of explicatures. Since thoughts are fully propositional,
semantic, conceptual and truth-conditional, Fodor’s (1998) requirement of semantic
compositionality, according to Carston no longer valid for language meaning in her
pragmatics, seems to remain necessary for the language of thought. This aspect has not been
clarified in the presented account.
To support this claim, Carston (2013, 198-199) discusses Wilson’s (2011) proposal that all words encode a
procedural meaning while open classes also encode concepts, advocating her own view.
9
3. Truth-conditional content in Relevance Theory
In relating her view on standing word meaning to truth-conditional semantics, Carston
juxtaposes minimalist semantics (Borg, 2012; Bach, 1994, 2010) to her own view, with the
original RT interpretation (Sperber and Wilson, 1986[1995]; Wilson and Sperber, 2004)
surfacing only occasionally.
According to Bach’s version of effability thesis “[…] all our thoughts are explicitly
expressible, in which case for every thought there is at least one sentence that would express
it explicitly.” (Bach, 2010, 129)
Consequently, competent speakers can always say exactly what they mean. This is
possible on the assumption of minimalist, externalist referential semantics explicating word
and sentence standing meaning. In this view, natural language sentences have a truthconditional semantics, that is, well-formed declarative sentences semantically express
propositions. A hearer’s grasp of an utterance’s truth-conditions is achieved with minimal, if
any reliance on context. Words are context insensitive and their standing meaning makes a
direct contribution to truth-conditional content. “Word meanings are concepts and it is these
concepts which stand in relation to objects in the world. Words have their content given by
things in the world.” (Borg, 2012, 144) In this way, open-class items contribute a determinate
denotation to the truth-conditional content of sentences (Carston, 2013, 194-195).
In the canonical form of RT (Sperber and Wilson, 1986[1995]; Wilson and Sperber,
2004), truth-conditional semantics was “responsible” for the meaning of thoughts expressed
in explicatures and, consequently, for the meaning of the latter. Explicatures have constituted
primary component of speaker meaning, derived from minimal linguistic content (encoded
meaning explicable in terms of linguistic semantics) via pragmatic enrichment and adjustment
in a given context. The explicitness of so derived explicature meaning comes in degrees
(Carston, 2013, 177). Pragmatic enrichment and adjustment consist in pragmatic inference
based on premises coming from beliefs, intentions, encyclopaedic information (and, thus, not
only from associative mappings and concept activation). Only intended contextual
assumptions enter the interpretation process (Carston, 2013, 178). The automatic decoding
process results in the (tacit) recovery of logical form (LF)/semantic representation (SR) of a
sentence, which is a syntactically structured string of atomic (“primitive”) concepts.10
In Carston’s account, the effability thesis collapses: “There are always components of a
speaker’s meaning which the linguistic expressions she employs do not encode.” (Carston,
2013, 202) Full explicitness/encoding/expressibility of content is generally not achievable, if
not for any other reason than because of the underdeterminacy thesis (Carston, 2002a).11
Another reason is that linguistically given meaning is schematic and non-propositional. It
only provides evidence and/or constraint on propositional content speakers communicate
(Carston, 2013, 177). Word meaning is not a conceptual, denotational entity. Consequently,
sentences do not encode thoughts/propositions and their meaning is non-truth-conditional.12 If
sentences are not truth-conditional, only one requirement on word meaning is left: “that those
word meanings are constant whenever they occur in a linguistic representation.”13 (Carston,
2013, 188). In fact, no single level of meaning can serve as both sentence semantics and
speaker-meant primary meaning (Carston, 2013, 176-177). Robyn Carston’s pragmatics is
compatible with Recanati’s (1994, 2004) contextualism, in contradiction to semantic
minimalism and effability thesis (Carston, 2002b; Sperber and Wilson, 1986[1995], 2012).
Truth-conditional semantics does not play any role in the description of language
meaning. It is assumed to give us the meaning of explicatures, that is of the thoughts
expressed by the speaker. Carston does not explain how exactly, if at all, it may happen.
10
Concepts are accessed via word-concept mappings (Sperber and Wilson 1986[1995]).
“The meaning encoded in the linguistic expression type that a speaker utters inevitably underdetermines the
content that the communicator [conveys], not only her implicatures but also the propositional content she
communicates explicitly (‘explicature’)” (Carston 2010, 4, 2009, 2002a)
12
Consequently, the principle of compositionality does not apply to language (Carston, 2013, 187).
13
Carston (2013, 188) claims that this requirement can be met even if word meanings are underspecified and
non-conceptual but she does not explain how this could be achieved (see pages 14,15)
11
Speaker’s thoughts, that is the meaning of explicatures, are truth-conditional (propositional),
conceptual, semantic and determinate. How can it be possible if there is no way to get from
schematic, non-conceptual, non-truth-conditional language meaning to conceptual speaker
meaning? What is the status and nature of such ad hoc atomic concepts and how can they
contribute to the truth-conditional meaning of the thought expressed by the speaker, subject to
the principle of compositionality, if their denotations may be not determinable and they may
have no logical entries?
Carston claims that her account of standing word meaning is compatible with the
general assumptions of Relevance Theory. It seems, however, that some fundamental
questions should be answered first.
4. Robyn Carston’s non-conceptual word meaning and Relevance Theory
In Relevance Theory the process of understanding utterances triggered by the two
principles of relevance14 is claimed to follow the relevance-theoretic comprehension heuristic:
Relevance-theoretic comprehension procedure: a. Follow a path of least effort
in computing cognitive effects. Test interpretive hypotheses (disambiguation,
reference resolution, implicatures, etc.) in order of accessibility. b. Stop when
your expectations of relevance15 are satisfied (or abandoned). (Wilson and
Sperber, 2004, 613)
Subtasks in the overall comprehension process: a. Constructing an
appropriate hypothesis about explicit content (EXPLICATURES) via
decoding, disambiguation, reference resolution and other pragmatic
enrichment processes. b. Constructing an appropriate hypothesis about the
intended
contextual
assumptions
(IMPLICATED
PREMISES).
c.
“Cognitive Principle of Relevance: Human cognition tends to be geared to the maximization of relevance.”
(Wilson and Sperber, 2004, 610) “Communicative Principle of Relevance: Every ostensive stimulus conveys a
presumption of its own optimal relevance.” (Wilson and Sperber, 2004, 612).
15
Relevance of an input is defined as a) Other things being equal, the greater the positive cognitive effects
achieved by processing an input, the greater the relevance of the input to the individual at that time. b) Other
things being equal, the greater the processing effort expended, the lower the relevance of the input to the
individual at that time. (Wilson and Sperber, 2004, 609)
Presumption of optimal relevance: a) The ostensive stimulus is relevant enough to be worth the audience’s
processing effort. b) It is the most relevant one compatible with communicator’s abilities and preferences.
(Wilson and Sperber, 2004, 612)
14
Constructing an appropriate hypothesis about the intended contextual
implications (IMPLICATED CONCLUSIONS). (Wilson and Sperber, 2004,
615)
The changes introduced to Relevance Theory by Robyn Carston in her papers “Word
meaning and concept expressed” (2012) and “Word meaning, what is said and explicature”
(2013) concern the (a) subtask of the RT comprehension heuristic, that is, explicature
construction via decoding, disambiguation, relevance resolution and other pragmatic
enrichment processes.
In the canonical version of RT, explicature construction has been described in many
papers. On hearing a speaker’s utterance U, the hearer, driven by the principles of relevance
and guided by the relevance-theoretic comprehension procedure, tries to recover the
communicated thought, that is, the speaker’s meaning. The thought is the truth-conditional
meaning of the explicature of the utterance.16 Explicature is a development of a logical form
encoded by an utterance (Sperber and Wilson, 1986[1995, 82]). Logical forms, as already
mentioned, are structured sets of atomic concepts, potentially undergoing individual-relative
pragmatic adjustment (Carston, 2010).
“[…] the hearer takes the conceptual structure constructed by linguistic decoding
(logical form), […] he enriches this on the explicit level and complements it at the implicit
level (guided by the cognitive principle of relevance).” (Sperber and Wilson, 1998[2012, 39])
The automatic decoding process is the first subtask17 in
explicature construction,
comprising semantically/linguistically encoded meaning (Carston, 2009, 2010). The resulting
logical form, being a starting point for its development into explicature, is conceptual in
nature.
16
The content of explicatures, or rather thoughts expressed by them is truth-conditional in character (Carston,
1991, 2004).
17
Utterance comprehension consists in online processing; individual subtasks are not sequential, but parallel.
In RT, concepts are “enduring, elementary mental structures, capable of playing
different discriminatory or inferential roles on different occasions in an individual’s mental
life.” (Sperber and Wilson, 1998[2012, 13])
They are accessed via words trough three entries (addresses in memory): lexical, logical
and encyclopaedic. Logical entries for concepts contain inference rules constituting truthconditional semantic content of the concepts. The rules are similar in the outcome of their
application to meaning postulates. On the one hand, the meaning of a concept is given by
those content constitutive inference rules and, on the other, the concept’s meaning is its
denotation. This denotation is inherited by the word encoding the concept as its meaning. In
the case of open-class items, there is a simple mapping from lexical forms to mental atomic
concepts (Carston, 2010, 8-9). Concepts may be shared by interlocutors or they may be
individual-relative and potentially non-communicable. They are arrived at through mutual
pragmatic adjustment of explicatures and contextual implications (Carston, 2010, 10).
In the relevance-search driven process of free pragmatic enrichment, ad hoc concepts
(marked with *) can be constructed or retrieved (Carston, 2002b, 2010). As already
mentioned, they are “pragmatically derived, generally ineffable, non-lexicalized […] rough
indication to aid readers in understanding what we have in mind in particular cases.” (Carston,
2010, 13) The meaning of ad hoc concepts is their denotation. They may be relatively stable
and retrieved from interlocutors’ memory. Such stable ad hoc concepts could have logical
entries and could contribute to the truth-conditional content of explicatures. Truly occasionspecific ad hoc concepts are constructed on the spur of the moment and their denotation may
be not determinable. Consequently, they could not contribute to the truth-conditional content
of explicit meaning.
The meaning of an explicature of an utterance, that is, the meaning of the thought
expressed by the speaker in the utterance is the outcome of the utterance’s logical form
development. The result is truth-conditional and propositional in character. The explicature’s
propositional form is “a well-formed formula which (a) undergoes formal logical operations
determined by its structure and (b) is semantically complete in that it is capable of being true
or false.” (Carston, 1991, 49)
Formal truth-conditional semantics operates on propositional forms. If no definite
propositional form is recoverable, formal truth-conditional semantics does not have anything
to operate on. The truth-conditionality of explicature meaning is possible because of its
conceptual character. The concepts result from (equally conceptual) logical form
development. To secure truth-conditionality, the concepts, constituting explicature meaning,
should have determinable denotations and, preferably, logical entries providing grounding for
truth-conditions. In the canonical RT version, these conditions can be met.
In Robyn Carston’s pragmatics, the situation is radically different. The divergences
concern the nature of decoding, interpretation of logical form and its development, the
character of mental ad hoc concepts constituting explicatures and explicature meaning in
general. All of them stem from the main assumption of Robyn Carston’s pragmatics,
according to which words do not encode concepts (so no lexicalized concepts are allowed)
because the schematic invariant word meaning, which is not open to pragmatic modifications
and adjustments and the nature of which has not been explicated yet, is non-conceptual and
non-truth-conditional.
In RT, logical forms are structured strings of concepts, resulting from the automatic
process of decoding the verbal input (prototypically in one’s native tongue). In view of
Carston’s new understanding of standing word meaning, this interpretation of logical form is
inadmissible for lack of lexicalized concepts. It is difficult to predict what changes in RT
would have to be introduced before a different account of the decoding process and the
resulting logical form character compatible with Carston’s pragmatics is introduced.
With the changing character of logical form, the process of its development would also
have to be modified. No individual-relative modulations or adjustments of the decoded
concepts can take place as there are no concepts on which they could work. As Carston
herself (2013, 197, 195 note 20) admits the greatest problem is how to get from the invariant,
non-conceptual, non-truth-conditional standing word meaning (which has not yet become part
of logical form and is not claimed to surface in the decoding process) to individual-relative,
conceptual, truth-conditional meaning of ad hoc concepts constituting the meaning of
explicatures. Until this question is answered, no logical form development can be postulated.
Since the only operational definition of explicature in RT is that it is a result of logical form
development18, the existence of explicatures cannot be predicted, either. Disregarding for a
moment the question of how one can get an explicature of an utterance otherwise than in the
process of logical form development, we might speculate on the nature of explicature in the
situation when words do not encode concepts and the process of ad hoc concepts construction
is obligatory. Then, explicatures would be composed of such occasion-specific ad hoc
concepts, constructed on the basis of words invariant schematic standard meaning and
contextual premises, as guided by the principles of relevance, but not necessarily the (a)
subtask of the relevance-theoretic comprehension heuristic. The question then arises what
would be the meaning of ad hoc concepts so constructed.
In RT the meaning of a concept is its denotation and a set of meaning constitutive
inference rules (analytic implications similar in outcome to meaning postulates) accessible via
the concept’s logical entry activated by the word serving as a pointer to the concept. In the
case of ad hoc concepts in the RT lexical pragmatics, their denotation is a narrowed and/or
broadened denotation of a concept encoded by the word providing the evidential input to the
ad hoc concept’s derivation. In Carston’s pragmatics ad hoc concepts’ denotations cannot be
The external criterion of explicature identification is the availability principle, according to which “in deciding
whether the pragmatically determined aspect of utterance meaning is part of what is said, we should always try
to preserve our pre-theoretic intuitions on the matter.” (Carston and Hall, 2012, 14)
18
established in this way because there are no lexicalized encoded concepts, related to ad hoc
concepts, on whose denotations the latter would base. The denotations of such new ad hoc
concepts, if determinable at all, would not be intersubjective, for lack of grounding in
lexicalized concepts meaning. The denotational status of non-conceptual, invariant, schematic
word meaning has not been clarified. Consequently, the denotation, that is, the meaning of
such new ad hoc concepts is not easily (if at all) predictable and determinable. Similarly,
truth-conditional meaning constitutive inference rules are most probably not accessible,
either, for lack of logical entries to such concepts. Such occasion-specific ad hoc concepts
lack enforced stability,19 which result in no access to pragmatic inference rules correlated with
word meaning. So, the question about those new ad hoc concepts meaning and nature seems
to be open. Until this is clarified, there are no grounds to claim their contribution to the truthconditional content of explicature, the operational definition of which has collapsed anyway,
since explicature cannot be the result of conceptual logical form development if there is no
such entity. How explicatures could express determinate, conceptual, semantic, truthconditional and truth-evaluable propositions with no conceptual input to occasion-specific ad
hoc concepts construction, which results in their denotational and inferential meaning being
not determinate (and possibly not determinable) and consequently not truth-conditional is a
problem that needs further explanation in Carston’s pragmatics. Until then, it is difficult to
postulate compatibility between RT and the new proposal, concerning the (a) subtask of the
relevance-theoretic comprehension procedure. An analysis of an example illustrates the
difference.
Example20
Context: Two acquaintances (A and B) are talking about John’s professional skills.
A: Is John a good surgeon?
The predicted lack of stability of such ad hoc concepts in Carston’s pragmatics would deprive them of their
conceptual status in canonical RT interpretation of concepts.
20
Various examples of category extension/narrowing/change analysis, including the “butcher” case reappear in
numerous RT papers (Carston, 2010; Sperber and Wilson, 1998[2012]).
19
B: He is a butcher.
[John is a bad surgeon]
In the example, the version of B’s utterance in square brackets is an approximation to its
explicature as possibly recovered by A.
In RT terms, what makes A recover the explicit content of B’s utterance is A’s search
for relevance as predicted by the two principles of relevance. The recovery of explicature
follows the relevance-theoretic comprehension procedure. A constructs a hypothesis about
explicit content, that is, the explicature of the utterance, via decoding, disambiguation,
reference resolution and other pragmatic enrichment processes such as ad hoc concept
construction. The automatic decoding process results in logical form recovery, comprising
semantically/linguistically encoded meaning (Carston, 2009, 2010). Logical forms are
structured sets of atomic concepts, potentially undergoing individual-relative pragmatic
adjustment (Carston, 2010). In the example, the structure of the decoded logical form of the
string of words He is a butcher could be BUTCHER (John). Guided by his search for
relevance, A accesses a mental atomic concept BUTCHER via simple mapping from the
lexical form of the open-class item butcher. The accessed concept may undergo individualrelative pragmatic adjustment, depending on A’s BUTCHER related assumptions and
experience. Since the potentially recovered explicature of B’s utterance with the literal
interpretation of BUTCHER [John earns money working as a butcher] does not confirm the
expected relevance of B’s utterance, the word butcher apart from encoding the concept
BUTCHER may serve as a clue to A to construct an ad hoc concept BURCHER*, which is
arrived at through mutual pragmatic adjustment of potential explicature and contextual
implications. This allows A to construct an explicature of B’s utterance [John is a bad
surgeon], which satisfies A’s expectation of relevance, against what he assumes makes a good
surgeon, and he stops processing. C’s possible explicature of the same expression could be
[John is a good surgeon], if C’s view of surgeons is that they should be like butchers – tough,
decisive and insensitive.
In Robyn Carston’s pragmatics, this simulation of the comprehension process is
inadmissible, for lack of conceptual logical form BUTCHER (John), for lack of lexicalized
mental atomic concept BUTCHER (no word-concept mapping), for lack of the possibility of
individual-relative pragmatic adjustment of the encoded concept BUTCHER, as there is no
such concept. The ad hoc concept construction, which in RT hinges on the relatedness
between the encoded concept and the newly constructed ad hoc concept, is inadmissible,
either, as there is no encoded concept to which the ad hoc concept could relate. The analysis
confirms the incompatibility of the two accounts. For lack of detailed interpretations of
invariant, non-conceptual standing word meaning and of the meaning and status of postulated
ad hoc concepts, an attempt at a similar simulation of utterance understanding process in
terms of Carston’s pragmatics would be unjustifiably speculative.
5. Conclusions
Over the nearly three decades of Relevance Theory development, the presented
interpretations of standing word meaning have changed. However, until Robyn Carston’s
(2013, 2012) new proposal, they assumed a conceptual (at least partly) nature of open-class
items content. The claim that stable, invariant word meaning is non-conceptual, non-semantic,
non-truth-conditional and schematic seems to be incompatible with the (a) subtask of the
relevance-theoretic comprehension procedure, concerning constructing an appropriate
hypothesis of an explicit content of an utterance. The incompatibility concerns the
interpretation of logical form, logical form development, the nature of ad hoc concepts and
the meaning of explicature.
References
Bach, K. (1994). Semantic Slack: What is Said and More. S. Tsohatzidis (Ed.), Foundations
of Speech Act Theory: Philosophical and Linguistic Perspectives (267-291). London:
Routledge.
Bach, K. (2010). Impliciture vs Explicature: What’s the difference? B. Soria and E. Romero
(Eds.), Explicit Communication: Robyn Carston’s Pragmatics (126-137). Basingstoke:
Palgrave.
Borg, E. (2012). Pursuing Meaning. Oxford: Oxford University Press.
Bosch, P. (2009). Predicate Indexicality and Context Dependence. Utterance Interpretation
and Cognitive Models. Current Research in the Semantics/Pragmatics Interface, 20, 99126.
Carston, R. (1991). Implicature, explicature and truth-theoretic semantics. S. Davis (Ed.),
Pragmatics: A Reader, (33-51). Oxford: Oxford University Press.
Carston, R. (2002a). Thoughts and utterances. The pragmatics of explicit communication.
Oxford: Blackwell.
Carston, R. (2002b). Linguistic meaning, communicated meaning and cognitive pragmatics.
Mind and Language, 17, 127-148.
Carston, R. (2004). Explicature and semantics. S. Davis and B. Gellon (Eds.), Semantics: A
Reader (817-845). Oxford: Oxford University Press. Originally published in 2000 in
UCL Working Papers in Linguistics, 12, 1-44.
Carston, R. (2009). The explicit/implicit distinction in pragmatics and the limits of explicit
communication. International Review of Pragmatics, 1, 35-62.
Carston, R. (2010). Lexical pragmatics, ad hoc concepts and metaphor: A Relevance Theory
perspective. Italian Journal of Linguistics, 22(1), 157-180.
Carston, R. (2012). Word meaning and concept expressed. The Linguistic Review, 29, 607623.
Carston, R. (2013). Word Meaning, What is Said and Explicature. C. Penco and F.
Domaneschi (Eds.), What is Said and What is Not (175-203). Stanford, California: CSLI
publications.
Carston, R. and A. Hall (2012). Implicature and Explicature. H-J. Schmid and D. Geeraerts
(Eds.), Cognitive Pragmatics, vol.4 of Handbooks in Pragmatics (7-84). Berlin: Mouton
de Gruyter.
Fodor, J. A. (1998). Concepts. Where Cognitive Science Went Wrong. Oxford: Clarendon
Press.
Frisson, S. (2009). Semantic Underspecification in Language Processing. Language and
Linguistics Compass, 3, 111-127.
Mioduszewska, E. (in press). Ad hoc concepts, linguistically encoded meaning and explicit
content. Some remarks on relevance-theoretic perspective. In press.
Recanati, F. (1994). Contextualism and Anti-Contextualism in the Philosophy of Language. S.
Tsohatzidis (Ed.), Foundations of Speech Act Theory (156-66). London and New York:
Routledge.
Recanati, F. (2004). Literal Meaning. Cambridge: Cambridge University Press.
Sperber, D. and D. Wilson (1986[1995]) Relevance. Communication and cognition. Oxford:
Blackwell.
Sperber, D. and D. Wilson (1998[2012]). The mapping between the mental and the public
lexicon. Wilson, D. and D. Sperber (Eds.) Meaning and Relevance (31-47). Cambridge:
Cambridge University
Sperber, D. and D. Wilson (2008). A deflationary account of metaphor. R. Gibbs (Ed.), The
Cambridge Handbook of Metaphor and Thought (84-106). Cambridge: Cambridge
University Press.
Sperber, D. and D. Wilson (2012). Meaning and Relevance. Cambridge: Cambridge
University Press
Wilson, D. (2003). Relevance and lexical pragmatics. Italian Journal of Linguistics/Revista di
Linguistica, 15, 273-291.
Wilson, D. (2011). The Conceptual-Procedural Distinction: Past, Present and Future. V.
Escandell-Vidal, M. Leonetti and A. Ahern (Eds.), Procedural Meaning: Problems and
Perspectives (3-31). UK: Emerald Group Publishing.
Wilson, D. and R. Carston. (2007). A unitary approach to lexical pragmatics: relevance,
inference and ad hoc concepts. N. Burton-Roberts (Ed.), Pragmatics (230-260).
Basingstoke: Palgrave Macmillan.
Wilson, D. and D. Sperber (2004). Relevance theory. L. Horn and G. Ward (Eds.), Handbook
of pragmatics (607-632). Oxford: Blackwell.
Streszczenie
Interpretacje konwencjonalnego znaczenia słów (rzeczowników, czasowników i
przymiotników) zmieniały się wraz ze zmianami samej Teorii Relewancji (część opisowa).
Jednakże, do czasu najnowszej propozycji Robyn Carston (2013, 2012), interpretacje te
zakładały
pojęciowy
skonwencjonalizowane,
charakter
omawianego
pozakontekstowe
znaczenia.
znaczenie
Propozycja
leksykalne
jest
Carston,
że
apojęciowe,
asemantyczne, nieprawdofunkcjonalne i schematyczne wydaje się stać w sprzeczności z tą
częścią relewancyjnej procedury opisującej proces rozumienia wypowiedzi, która dotyczy
konstrukcji jej eksplikatury (część dyskursywna: krytyka propozycji Carston z perspektywy
Teorii Relewancji).
Summary
The Relevance Theory interpretations of standing word meaning have changed during
the evolution of the theory itself (the expository part of the paper). However, until Robyn
Carston’s (2013, 2012) new proposal, a conceptual (at least partly) nature of open-class items
content was assumed. The claim that stable, invariant word meaning is non-conceptual, nonsemantic, non-truth-conditional and schematic seems to be incompatible with the (a) subtask
of the relevance-theoretic comprehension procedure concerning constructing an appropriate
hypothesis about explicit content of an utterance (the discursive, critical evaluation of
Carston’s proposal).
Download